Production Coelomomyces, FungusThat Kills Mosquitoes › content › pnas › 69 › 8 › 2043.full.pdf · southeastern United States and after two trips to India, I finaly found

Proc. Nat. Acad. Sci. USAVol. 69, No. 8, pp. 2043-2047, August 1972

Mass Production of Coelomomyces, a Fungus That Kills Mosquitoes(mosquito larvae/fungus/sporangia)

J. N. COUCH

Department of Botany, University of North Carolina, Chapel Hill, N.C. 27514

Contributed by John N. Couch, May 26, 1972

ABSTRACT In work on the control of mosquitoes bythe fungus Coelomomyces, the main problem is a sourceof inoculum since the fungus has not been culturedartificially with production of sporangia. We reared thelarvae of Anopheles quadrimaculatus in algal water insteadof in water with soil. By addition of inoculum once ortwice in small amounts, the larvae become infected, andmany grow to large fourth instars whose bodies are filledwith sporangia. Such larvae are perfect for inoculum. Ifinoculum is added in much larger amounts and so timedthat sporangia will be discharging spores during the first,second, and third ecdyses up to 100%, infection occurs,most of the larvae dying as late second or early third in-stars. This type of infection is good for extermination ofmosquitoes but not for production of inoculum. Crudefield tests have averaged 60% infection.

In spite of the wonders performed by chemical insecticides,mosquitoes still transmit many of man's worst diseases. Infact, in many parts of the world, malaria is on the increasebecause the mosquitoes that transmit it have become resistantto 1,1,1-trichloro-2,2-bis(p-chlorophenyl)ethane (DDT), andmore potent insecticides are needed. It is essential, therefore,that other methods of controlling mosquitoes be found tosupplement the chemical insecticides.As a consequence, several predators and parasites of

mosquitoes are being studied as a means of control. Amongthe parasites, the most promising is the fungal genus Coelo-momyces first described by Keilin in 1921 in a larva of theyellow fever mosquito from near Singapore (1). Since thenit has been reported from every continent and ranges fromthe equator to central Alaska. About 30 species have beendescribed, some of which attack several of the major disease-transmitting mosquitoes. The species are obligate parasites,each having as a rule its own particular mosquito host, andunlike other fungal parasites, they do not attack other insectsor other animals. The fungus enters the coelom of the livinglarva and grows at the expense of the fat body. In the case of aheavy infection, the fat body is consumed and the larva fails topupate since the pupa depends on the food stored in the fatbody for its nourishment. The mycelium or vegetative part isunique in all the true fungi in that it lacks a rigid cell wall,the absence of which is a fitting adaptation for life in thehemocoel of a wiggler. The mycelium at maturity is trans-formed entirely into sporangia, and these, under favorable con-ditions, germinate producing many minute swimming sporesthat escape and infect a new crop of larvae.A fungus such as Coelomomyces, fatal to mosquitoes and

harmless to other organisms, would seem ideal for biologicalcontrol. However, since it has not been cultured outside its

host, one must either collect parasitized larvae in the field orrear the fungus in its larval hosts in order to get a supply ofinoculum. Walker (2), Muspratt (3), and Madelin (4) haveall succeeded in infecting healthy larvae, but only Madelin hasbeen able to carry on the infection for any length of time, andall three were successful only when soil from the locationwhere the infection occurred was present in the containers.Madelin states that only a small proportion of the infectiontests were successful.

I have carried on the infection of Coelomomyces in a mos-quito without the use of soil, and by new methods of handlingthe inoculum, I have been able to greatly increase the rate ofinfection and the reliability of results. This preliminarypaper reports these advances.

MATERIALS AND GENERAL METHODS

Experimental work on infection would be impossible withouta readily available supply of host and parasite material. Themosquito should be easily reared in the insectary and itsstructure and life history should be well known. For precisework, it is essential that the sporangia of the parasite can begerminated. It also adds zest to the work if the mosquito is avector in a human disease.

After searching for experimental material throughout thesoutheastern United States and after two trips to India, Ifinaly found ideal material for studies of infection in ourlocal water supply, a lake about 3 miles from the Universitycampus. The host was our common malarial mosquito,Anopheles quadrimaculatus; the parasite was Coelomomycespunctatus. Larvae occurred in shady coves, in shallowwater where there was aquatic vegetation and floating planttrash, giving them protection from small fish. Parasitizedlarvae can be recognized by the reddish-brown color of thethorax and abdomen, the color being in the walls of mature,closely packed sporangia. The infection was found in Uni-versity Lake September 24, 1965, and from this date until theend of October, frequent trips were made to the lake to collectparasitized larvae for inoculum. These were placed in rows ondamp filter paper in petri dishes, about 50 larvae to eachdish, and stored in a refrigerator at 100. If the filter paper iskept moist, the sporangia remain capable of germinating forat least 5 months.A unique advantage of the Coelomomyces punctatus-

Anopheles quadrimaculatus combination is that severalstrains of this mosquito have been available to us through thekindness of the Communicable Disease Center at Savannah,Ga. In the infection studies we have used Savannah, Cleve-

2043

Dow

nloa

ded

by g

uest

on

July

23,

202

0

Proc. Nat. Acad. Sci. USA 69 (1972)

FIG. 1. Head, thorax, and two segments of abdomen almostcompletely filled with oval, mature sporangia of Coelormomyces.Normal infection; produces inoculum. About X 75. A composite offour separate photographs by C. Bland.

land, Bethesda, and Hartwell strains from Savannah, all ofwhich are susceptible to the fungus.The eggs of An. quadrimaculatus are not ready to hatch

immediately after being laid, but require an incubationperiod that is accelerated by heat and delayed by cold. Theoptimum temperature for maximum hatching is near 330 andrequires about 33 hr (Horsfall, ref. 5). In order to have a con-

stant supply of eggs we keep them in closed containers on

damp filter paper at 100. Under such conditions the eggs re-

main capable of hatching for about a week. When the larvaeare 24- to 36-hr old, they are fed formula "A" furnished us bythe Communicable Disease Center at Savannah, Ga.

Infection experiments are performed in a section of thegreenhouse, since this location more nearly simulates naturalconditions. During summer the roof and glass sides are keptpainted with a white greenhouse paint, which, with the aid of

an evaporative cooler, keeps the temperature approximatelybetween 18 and 35°. In winter the temperature varies fromabout 25 to 30°.We have used a variety of containers for the infection

experiments: small rectangular pans 25 X 40 X 7 cm deep,large white enamel and plastic dishpans, and large rectangularsinks 40 X 74 X 10 cm deep, in all of which we have obtainedinfection. Infection has occurred in lake, tap, rain, and snowwater, with the best rate in lake water.

DETAILED METHODS AND RESULTSSporangial germinationI had observed stages in the germination of the sporangia in1947 but, because of the small supply of living material, wasunable to follow the process in detail. With abundant material,it was possible to work out a proper schedule and to photo-graph the main features of the process. For observations ongermination, a heavily infected larva (Fig. 1) is placed on aclean slide in a large drop of water, and the body is openedwith dissecting needles so that the sporangia are exposed tothe water. Such a larva may contain over 60,000 sporangia.The slide is placed in a damp chamber prepared by put-ting two round pieces of filter paper in the bottom of a petridish with a V-shaped glass tube 5 mm in diameter on thepaper to support the slide. Enough water is poured into thedish to wet the paper, and thus prevent drying of the spor-angia. Material so prepared should remain moist for 48-72 hrwithout addition of water, permitting easy and frequentexamination of the sporangia.The time schedule for the germination of the sporangia is

controlled by environmental factors and varies according tothe treatment the sporangia have received (6). The larvaemay be taken from their natural habitat or reared in thegreenhouse, alive or not long dead, or stored for several weekson moist filter paper in the 100 refrigerator. It is best that thelarva's cuticle be intact, otherwise fresh water may haveentered the coelom and caused the sporangia to germinate. Inmost cases the damp chambers containing the sporangia wereplaced on my laboratory table about 122 cm (4 ft) from alarge west window, and thus exposed to diffuse daylight andsome afternoon sun. If fresh material is used, i.e., sporangiathat have not become dormant, germination occurs equallywell on the laboratory table in alternating daylight anddarkness and in a completely dark incubator. Althoughgermination appears not to be influenced by light whennewly matured sporangia are used, further experimentation onthis matter is needed. Temperature has a marked effect ongermination; no germination occurrs below 100 or above350, the optimum being between 21 and 270. Germinationoccurs in tap, distilled, rain, and lake water, in pure cleanwater, and in water contaminated with bacteria and protozoa,but not in water containing the waste products of manylarvae.

Germination takes about 48 hr at a temperature of around230 and has been followed under the microscope many-timeswith attention to changes in the organization of the,protoplasm (6). The more conspicuous features of germinationare as follows. A mature sporangium shows an amorphouscentral mass of small lipoid bodies. In germinating thiscluster becomes dispersed during the first 24 hr, the contentsnow showing a sort of alveolate arrangement. About 6 hrbefore the sporangium discharges, a large lateral bulge

OTUA Botany: Couch

Dow

nloa

ded

by g

uest

on

July

23,

202

0

Inoculum for Infection of Mosquito Larvae 2045

pushes out causing the outer wall to open at a preformed slit.At this time the lipoid bodies are of about uniform size andare dispersed in the cytoplasm. After about 48 hr, when thesporangium is ready to discharge its spores, the lipoid bodiesseen under high dry are so grouped in each spore that thecluster viewed edgewise has a rod-like appearance. Sporangiain this "go" condition are ready to discharge their spores, andif covered with a cover glass, will empty in about 15 min. Thatthe emergence of the zoospores under the cover is not causedby the pressure of the cover glass has been demonstrated byputting the slide in an anaerobic chamber from which oxygenhas been largely removed by a mixture of pyrogallic acid andsodium hydroxide. In such a chamber the sporangia in the"go" stage discharge after 10-20 min (6). While such treat-ment hastens sporangial discharge, it is not necessary.Given the proper environmental conditions with the tempera-ture 21-27° most of the sporangia of C. punctatus germinate,maturing and discharging their spores in a period of about48 hr. A few sporangia may discharge spores in less than 48 hr,a few may discharge up to 72 hr or longer, and a fewmay remain unchanged with a central mass of lipoid bodies,apparently in a dormant stage. As a rule, however, if newlymatured sporangia are used the great mass of sporangia,50-60 thousand from one larva, discharge their spores afterabout 48 hr. In fact on such slides, the mass of swimmingspores is so great that the water has a distinct milky appearanceon the side of the drop nearest the window, since the zoosporesare phototactic.A disruption in the 48-hr schedule may occur when infected

larvae are stored on damp filter paper at 100 for 4 months orlonger and permitted to dry once or twice during this period.Such slides show a continuous discharging of a few sporangiaat a time over a period of about 10 days. The 48-hr schedule isperhaps followed in nature during the warm summer monthsand the slower one follows the winter months of freezing anddrying. The mosquitoes appear in University Lake aboutmid-April, but so far larvae infected with Coelomomyces havenot been found until after mid-May.Other workers who have obtained laboratory infection have

used soil in their containers from a location where infectedlarvae occurred, and until mid-summer 1968, I thought thatsoil was necessary (6). With soil, results were erratic, usuallyless than 50% of the containers having infected larvae.However, if no soil was added to the water, as a rule, no infec-tion occurred.Because of the disadvantages of soil, in early summer of

1968 I began a study to eliminate its use. While this was inprogress I continued to use soil from University Lake in orderto avoid losing the infection. Up to this time, I had suspectedthat the presence of a heavy growth of algae in the water pre-vented infection since I had found that the sporangia in the"go" stage germinated more rapidly if put in an atmospheredevoid of oxygen. In one routine experiment with 12 panscontaining soil with aerated tap-water, each with inoculumand eggs prepared in the usual way, one of the pans hadlarge, plump, heavily infected larvae. The bottom and sidesof the pan were covered with a dense growth of unicellulargreen algae. The commonest of these were several speciesof Scenedesmus, Ankistodesmus falcatus, and a small spe-cies of Closterium. When some of these infected larvaewere dissected, the guts were seen to be filled with the

isolated them in pure culture and kept them going on Keilin-Das agar in flasks (7). The great advantage of the algae is thatwe get a consistently high rate of infection without soil.The larvae in feeding take in the small unicellular algae alongwith formula "A," the algae passing- out in large quantitiesthrough the gut and anus with other fecal matter. The algaemultiply rapidly in the water rich with the nitrogenous wastesfrom the larvae. In pans with a good growth of algae, thelarvae are larger and the water is clean and free from the un-pleasant odor of pans in-the insectary where larvae arefed on formula "A" alone. The larvae use the algae as food,and the algae use urea and other waste products given off bythe larvae. Furthermore, I have found that the sporangia ofthe Coelomomyces fail to germinate in small vessels containingmany larvae if no algal growth is present.

After having established the value of the algae in increasingthe amounts of infection, we are now trying to improve ourmethods of handling the inoculum. To keep the work goingwe need a large supply of inoculum. This is obtained byharvesting infected larvae from successful experiments asfollows.- A batch of about 50 heavily-infected larvae are takenfrom the pans and sinks and put in a quart jar half filled withalgal water, covered loosely, and stored at 100 until needed.The inoculum is prepared for use by pouring the contents of

a jar through a very fine steel mesh. The water, algae, andany loose sporangia of Coelomomyces pass through the mesh.The larvae that are caught on the sieve are pressed throughwith the fingers or with a small pestle; this process separatessporangia from broken parts of larvae. Good inoculum shouldshow 5-10 sporangia in each small drop under low powerof the microscope. 500 ml of such inoculum can-be diluted to2000 ml, and this would be enough-inoculum and algal waterto use in 10 sinks or dishpans or 20 small pans.For infection experiments, the containers are filled to a depth

of about 5 cm (2 inches) with lake or aerated tap water. 100-200 ml of the inoculum are added to each container. A squarepiece of wax paper with the center cut out, making an openingabout 2 cm2, is placed on top of the water. The eggs (200-800) on a strip of filter paper are dropped through theopening; the eggs remain on the surface of the water sur-rounded by the wax paper and the filter paper sinks to thebottom. The algae should be so abundant as to give -the waterin the sink or pan a distinct greenish tint. By using algalwater mixed with inoculum, we can be sure that all pans willhave some to many infected larvae.

Several experiments have been run to determine wheninfection occurs by putting inoculum with the first, second,third, and fourth instars. The results indicate that infectionmay occur during any of the stages. Such experiments sug-gested a completely new method of applying the inoculum.If infection occurs during any one of the four instars, whatwould be the result if not one dose of inoculum but fourdoses were given, the time of application so scheduled thatgerminating sporangia would be releasing spores when ecdysiswas occurring in the first, second, third, and fourth instars?Such an experiment was started on May 9, 1970 in 12 pans

each with lake water. 200 ml of inoculum and about 300 eggsof An. quadrimaculatus were added to each pan. The eggswere 3 days old and began hatching immediately. Moreinoculurn was added May 11, 13, 16, and 19. The larvae

same algae. In order to maintain a supply of these algae I

Proc. Nat. Acad. Sci. USA 69 (1972)

were fed formula "A" twice daily.

Dow

nloa

ded

by g

uest

on

July

23,

202

0

Proc. Nat. Acad. Sci. USA 69 (1972)

On May 24, the larvae were examined for the presence ofCoelomomyces, either mycelium or sporangia. For an ac-

curate count of the infected larvae and pupae in a pan,

the entire contents were poured into a container with a finestainless steel mesh bottom. The water, algae, and protozoapassed through the mesh leaving the larvae and pupae behind.These were then washed under the faucet and poured into a

large finger-bowl, where in clean water they could be betterseen. The heavily infected fourth instars with mature sporangiacould easily be picked out because of their reddish-browncolor. Less heavily infected larvae were examined underlow power of the microscope for the presence of mature andimmature sporangia, hyphal bodies, and mycelium. When thelarvae reached the fourth instar stage, they were carefullyand frequently observed, for even when the rate of infectionis very heavy, some larvae may escape infection or be lightlyinfected, pupate, and then escape as adults. In order tohave an accurate count such pupae must be caught andexamined for the presence of fungus.

1714 larvae were taken from the pans and all were infected.Of the 53 pupae recovered, 50 were infected. This experimenthas been repeated many times with similar results.

Since finding that the infection rate is close to 100% whenthe inoculum is added to the larvae at intervals so spacedthat germinating sporangia will be present during each ofthe four ecdyses, I followed this schedule until by chanceI discovered the most susceptible stage of the larvae to infec-tion by Coelomomyces.

In a recent experiment because of the short supply, inocu-lum was added only twice, first with the eggs and then againto the first instars 2 days later. Six large dishpans were

used, each started with 800-1000 eggs from which about600 instars survived in each pan. On the tenth day I noticedmany larvae clustered at the surface of the water against theside of one of the pans, a certain sign of sick larvae. I ex-

FIG. 2. Part of anal segment, with evaginated coelom. contain-

ing vast numbers of spherical hyphal bodies. Abnormal infection;

amined 12 from the cluster, and the coelom of each wasfilled with spherical hyphal bodies of fungus. 5 Dayslater 3251 larvae were examined; 3211 were parasitized, aninfection rate of 98.7%. Because of the surprising results, theexperiment was repeated on a larger scale; 10 dishpans and10,000 eggs were used. Infected second instars were recognizedafter 7 days. 3 Days later, many of the larvae had collected inmotionless groups. The larvae in 2 of the 10 pans were ex-amined, and 1244 were infected. No uninfected larvae wereseen in any of the pans. The vast majority was so heavilyinfected that they died as late second or early third instars.In such larvae the fungus usually fails to form sporangia anddisintegrates as the larvae die. The larvae less heavily in-fected reached the third and fourth instar stage and con-tained mature sporangia. No pupae were observed in any ofthe pans. Such a heavy infection resulting in exterminationof larvae in the greenhouse is an abnormal condition rarelyoccurring in nature (Fig. 2). Only Muspratt (8) has reportedsuch a high rate of infection.At a temperature of about 270, this overkill shows up when

the larvae are between 7 and 10 days old. The sick larvaecollect in groups and continue to feed and remain active for afew hours, then they drift to the sides of the pan, and in a fewhours, die and sink to the bottom. If such larvae are examinedunder the microscope while still active, the coelom is seen tobe filled with spherical or irregular hyphal bodies (Fig. 2).There is little or no mycelium and no signs of any myceliumattached to the fat body. The vast number of hyphal bodiesfloats freely in coelomic fluid, receiving their nourishmentfrom it, and thus they cause the death of the larvae, perhapsfrom starvation. When the sick larvae become inactive anddeath sets in, the hyphal bodies begin to swell, burst, anddisappear completely. Rarely a few sporangia may havematured far enough to form the thick brownish wall beforethe larva began to die. Such sporangia persist in deadlarvae, indicating the former presence of heavy infection.This overkill type of infection that has not been observed byany previous worker, is useful if one wishes to exterminate apopulation of larvae but is practically worthless in theproduction of inoculum.For the production of good inoculum, i.e., large fourth

instar larvae whose body cavities are full of the sporangia ofCoelomomyces, much less inoculum is required than to produceoverkill, and it may be added once or twice with the eggs orwith the instars. Such treatment usually produces a typicalinfection as found in nature, in which the hyphal bodies becomeattached to the fat body at various places throughout thecoelom and develop into mycelia that grow and branch,finally consuming the fat body. In such infection, very fewhyphal bodies or hyphagens are seen. As the infection matures,the mycelia are completely used up in the formation ofsporangia (Fig. 1). By the methods described above, it isnow possible to produce unlimited quantities of inoculum ofCoelomomyces punctatus, given the necessary space, equip-ment, and personnel.

NATURAL MEANS OF DISPERSALOF COELOMOMYCES

Among the agents of distribution of the fungus are wind,birds, and fish, but most important are adult mosquitoes.Lightly infected mosquitoes may emerge from the pupal caseand fly off carrying mature sporangia in their coeloms.

2046 Botany: Couch

all such larvae die, as does the immature fungus. About X 180.

Dow

nloa

ded

by g

uest

on

July

23,

202

0

Inoculum for Infection of Mosquito Larvae 2047

Heavily infected adults may fly along the water surfacevainly pulling their pupal cases behind like a water skier but,because of the weight of the sporangia, they are unable to takeoff.

RESULTS OF FIELD TESTS

To test the effectiveness of C. punctatus in controlling thebreeding of Anopheles quadrimaculatus in the field, we haveused ditches in land made available at the sewage disposalplant. The tests have been conducted through four summersbeginning in 1968. After May 15, eggs of An. quadrimaculatuswere put in the ditches with inoculum in algal water, and10-15 days later collected and examined for the fungus.Depending on the availability of eggs and inoculum, thetests were repeated until frost in October. The infection ratevaried from zero to near 100%, the average being 60%.With an adequate supply of inoculum, eggs, and help, we areplanning more definite tests this summer, 1972, using theabove mentioned ditches and the natural breeding sites in theUniversity Lake.

In the early stages of this work I was helped by J. Roan, C. J.Umphlett, C. Bland, W. Martin, P. Lum, Elmo McCray, andDon Ashton. The mosquito eggs were supplied by the Com-

municable Disease Center, PHS at Savannah, Ga. Supportedby NIH Grant AI-03235.

1. Keilin, D. (1921) "On a new type of fungus: Coelomomycesstegomyiae, n.g., n.sp., parasitic in the body cavity of the larvaof Stegomyia scutellaris Walker (Diptera, Nematocera, Culi-cidae)," Parasitology 13, 22S-234.

2. Walker, A. J. (1938) "Fungal infections of mosquitoes,especially of Anopheles costalis," Ann. Trop. Med. Parasitol.32, 231-244.

3. Muspratt, J. (1946) "Experimental infection of the larvae ofAnopheles gambiae (Dipt., Culicidae) with a Coelomomycesfungus," Nature 158, 202.

4. Madelin, M. F. (1968) "Studies on the infection by Coelomo-myces indicus of Anopheles gambiae," J. Elisha Mitchell Sci.Soc. 84, 115-124.

5. Horsfall, W. R. (1955) in Mosquitoes, Their Bionomics andRelation to Disease (The Ronald Press Company, New York),p. 135.

6. Couch, J. N. (1968) in "Sporangial germination of Coelomo-myces punctatus and the conditions favoring the infection ofAnopheles quadrimaculatus under laboratory conditions,"Proceedings of the Joint U.S.-Japan Seminar on MicrobialControl of Insect Pests (Fukuoka, Japan), pp. 93-105.

7. Das, G. (1968) "Growth and appearance of Scenedesmus asinfluenced by inorganic nutrition," Sv. Bot. Tid~kr. 62, 459.

8. Muspratt, J. (1963) "Destruction of the larvae of Anophelesgambiae Giles by a Coelomomyces fungus," Bull. W. H. 0.29, 81-86.

Proc. Nat. Acad. Sci. USA 69 (1972)

Dow

nloa

ded

by g

uest

on

July

23,

202

0